Smart Trap Detecting the Whitefly Pest using Deep Neural Networks

This study analyzes the suitability of using convolutional neural network designs for detecting the whitefly pest in different agricultural contexts. The goal of this research work is to develop a smart trap suitable for field use. As an initial step, the existing state-of-the-art whitefly detection and counting techniques are reviewed. Then, the datasets related to whitefly infestations in diverse environments, encompassing both laboratory and field conditions are collected. The proposed approach involves training with various algorithms tailored for object detection with the collected datasets. Subsequently, the results are evaluated and compared to identify the most effective architectures.